The BUQO algorithm provides standalone functionality for the quantification of uncertainty quantification of local structures appearing in images formed as Maximum A Posteriori (MAP) estimates, assuming that the underpinning Bayesian model is log-concave. It can typically be used as a complement to algorithms of the SARA family, which are not shipped with direct uncertainty quantification functionality. It finds application in establishing the confidence in detected lesions in medical imaging, astrophysical sources in astronomical imaging, etc. BUQO's high scalability in comparison to sampling approaches stems from the fact that it is propelled only by optimisation algorithms.
